Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Theoretical bounds of majority voting performance for a binary classification problem.

Anand Narasimhamurthy1

  • 1Department of Computer Science and Engineering, 341 IST Building, Pennsylvania State University, University Park, PA 16802, USA. narasimh@cse.psu.edu

IEEE Transactions on Pattern Analysis and Machine Intelligence
|December 17, 2005
PubMed
Summary

This study presents a new optimization framework for majority voting in binary classification, removing independence assumptions. It establishes theoretical performance bounds without needing classifier independence, enhancing ensemble methods.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Automated interventricular septum thickness measurement from B-mode echocardiograms.

Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Interventionยท2010
See all related articles

Area of Science:

  • Machine Learning
  • Computer Science
  • Data Science

Background:

  • Previous theoretical analyses of majority voting relied on classifier independence.
  • This assumption limits the applicability of existing models in real-world scenarios.

Purpose of the Study:

  • To develop a theoretical framework for majority voting that does not assume classifier independence.
  • To establish performance bounds for majority voting by formulating it as an optimization problem.

Main Methods:

  • Formulated majority voting as an optimization problem with linear constraints.
  • Analyzed the objective function, identifying linear programming (LP) cases and nonlinear cases with an even number of classifiers and rejection.
  • Derived theoretical upper and lower bounds for performance.

Related Experiment Videos

Main Results:

  • The theoretical performance bounds for majority voting are solutions to the formulated optimization problem.
  • The framework accommodates scenarios without classifier independence.
  • Investigated the relationship between classifier diversity measures and majority voting performance.

Conclusions:

  • The proposed optimization framework provides a robust method for analyzing majority voting performance.
  • It offers insights into ensemble methods by removing restrictive independence assumptions.
  • The study advances the theoretical understanding of classifier combinations.